Choose SKU mapping software that connects an exact supplier product or variant to the correct selling record, exposes ambiguous matches and preserves the reason for each relationship. Test duplicates, pack sizes and changed identifiers before allowing the mapping to support stock decisions. A similar product title is not enough evidence.
A mapping error can look surprisingly tidy. The source check succeeds, the software calculates a quantity, and an operator sees a normal-looking proposal. The problem is that the observation belongs to another size, colour or pack. Good mapping controls make that difference visible before the proposed change reaches a live listing.
This guide helps supplier-backed eBay sellers compare mapping workflows. It focuses on buying criteria and evaluation tests. Zelluvo is the publisher and offers supplier-to-listing mapping as part of its documented workflow. The examples below are invented records for explanation, not screenshots, customer data or measured product results.
Understand the records a mapping connects
A supplier SKU identifies a record within the supplier’s system. Your internal SKU identifies a record within your own operation. An eBay listing identifier identifies a marketplace listing. A listing may also contain individual variations, each with its own identifying information. Software needs enough context to join the right records without confusing their roles.
Do not assume that identical text makes two codes globally equivalent. Two suppliers might both use “1001” for unrelated products. Your matching key may therefore need the supplier identity as well as the supplier SKU. Likewise, the account and listing context matter when resolving a marketplace record.
A GTIN is a different kind of identifier. GS1 describes it as a global product identifier in its explanation of GTINs and barcodes. Do not rename a supplier’s internal code as a GTIN merely to populate an empty field. Preserve the meaning and provenance of each identifier.
| Field | Example | Purpose |
|---|---|---|
| Supplier | Example Wholesale A | Disambiguates the source catalogue |
| Supplier variant SKU | TEE-BL-M | Identifies blue, medium at the supplier |
| Internal SKU | APP-104-BL-M | Links the business’s own record |
| Marketplace record | Test listing / blue / medium | Identifies the intended destination |
| Pack quantity | 1 | Prevents a unit being confused with a multipack |
| Review state | Verified in test | Separates reviewed relationships from suggestions |
Separate suggested matches from accepted relationships
Automatic suggestions can reduce manual work, but a suggestion should remain distinguishable from a verified mapping. Ask which fields support the proposed match. An exact identifier with consistent attributes is a different level of evidence from a similar title and photograph.
During a demonstration, provide two products whose names differ only by a size or pack count. Then inspect what the software suggests and what a reviewer can see. If it hides the distinguishing attribute, a high match rate is not a useful buying signal. The system needs to expose uncertainty, even if that means more rows remain unresolved.
Check whether the tool can leave a record unmapped. Forcing every row into a match creates the appearance of completion at the expense of accuracy. A useful exception list tells the operator which attribute or identifier is missing and what evidence would resolve it.
Also ask how an accepted mapping is changed. A later import should not silently replace a reviewed relationship merely because a new title looks more similar. Whether the product uses manual controls or automated rules, the change should be explainable and visible to the people who depend on it.
Test the awkward cases in your catalogue
Colour and size
Use two variants that share the same parent product. Confirm that the mapping stores the distinguishing options and observes the correct source selection. A parent-level product link may be useful for context, but it does not establish that every child variant has the same stock or cost.
Single items and packs
Compare a single unit with a two-pack. A source quantity of ten packs is not automatically ten saleable individual units, nor should it become twenty without a documented conversion. If your business breaks packs, establish how the tool represents that relationship and who maintains the rule.
Repeated identifiers
Introduce a duplicate supplier code in a different catalogue or source account. Ask whether the software identifies the ambiguity, uses sufficient context or simply picks a row. The desired behaviour is an explicit, defensible relationship rather than a silent preference.
Discontinued and replacement products
A supplier may replace an item with a visually similar model. That is not necessarily the same product. Review compatibility, attributes and listing promises before accepting a replacement. A tool should not treat a changed source page as proof that the original goods remain available.
Formatting differences
Leading zeros, spaces and letter case can change the way identifiers survive imports. Decide which fields are text and preserve their original values. Any normalisation should be deliberate and reversible. Removing a leading zero because a spreadsheet interpreted a code as a number can create a new mismatch.
Ask how mappings affect stock proposals
A mapping is a relationship, not an instruction to copy every source value. Establish what the software does after it resolves the source. Does it create a local observation, calculate a proposed quantity, or attempt a live update? Those steps have different consequences and should have different controls.
Review the current destination quantity and the source evidence together. If a supplier variant is available but the mapping was only suggested, a live change should not be treated as routine verified work. The operator should know which part of the chain is still uncertain.
For eBay implementations, the developer documentation on SKU identifiers explains that SKU behaviour depends on listing and tracking context. Do not assume one identifier rule applies identically to every listing method. Your software provider should be able to explain which record it uses for your account’s workflow.
Zelluvo’s documented design keeps live eBay stock changes approval-first. Supplier mapping and local proposals support the review; the connection itself is not permission to change a listing. When evaluating it, confirm the product, variant and destination shown before approval, then verify the result separately.
Plan for mapping changes after setup
Initial matching is only part of the work. Suppliers rename products, discontinue variants and alter catalogue structures. Your own listings can also change. Ask how the system surfaces a relationship that no longer appears valid and how an operator can investigate it without losing the original evidence.
A useful change record includes the previous source, the new source, the affected destination and the reason for the correction. If a pending stock proposal depends on the old relationship, reassess it before approval. Otherwise, correcting the mapping can leave an outdated action waiting in the queue.
Specify who is allowed to edit mappings. In a small team, that may be the same person who reviews stock, but the responsibility should still be explicit. If several people import files, agree how they avoid overwriting each other’s work or recreating a previously rejected match.
Review a sample periodically and after material changes. A random sample can reveal formatting or maintenance problems; a targeted sample can focus on troublesome suppliers and variants. Record what was checked and why. Avoid claiming the entire catalogue is accurate because a small sample passed.
Evaluate imports, exports and recovery
Check access boundaries as part of the trial. A reviewer should see and change the mappings belonging to the intended account or organisation. If you support several businesses, a convenient combined view must not blur ownership. Ask the provider to demonstrate account selection and how a mistaken cross-account destination is prevented. This is particularly important when supplier codes look identical across separate catalogues.
Bring a sanitised mapping file to the trial. Include the columns your business actually uses, plus a few deliberate exceptions. Check whether the importer previews changes, explains rejected rows and preserves the identifiers exactly. A clear row-level error is more useful than a general “import failed” message.
Find out what a second import does. Does it create duplicates, update existing relationships or leave them unchanged? Use a small test set where you can inspect every result. Repeatability matters because supplier files and team processes rarely arrive perfectly once and never change again.
Ask what information you can export or otherwise retain. At minimum, your operating records should remain understandable if you change tools. Confirm the available fields and format rather than assuming an export button includes source context, variant attributes and review history.
Recovery should be part of the demonstration. Introduce one incorrect mapping in a safe test record, correct it, and inspect the resulting history and dependent proposals. Do not deliberately create a wrong live listing update to test recovery. The objective is to prove that the correction process is controlled before real stock depends on it.
Run a mapping evaluation with clear acceptance tests
Define success before the sales call. A small evaluation might contain one simple product, two close variants, one duplicate code, one missing identifier and one pack conversion. These are suggested categories, not a statistically representative sample. Adjust them to the failure patterns in your own catalogue.
- Write the intended source and destination for every test record.
- Capture the attributes that distinguish otherwise similar products.
- Import or create the records without permitting live stock changes.
- Inspect suggested matches and leave ambiguous cases unresolved.
- Accept only relationships supported by the source evidence.
- Generate or inspect a local proposal where the product supports that workflow.
- Change one mapping and check how dependent work is handled.
- Export the results or retain an equivalent review record.
Keep the evaluation notes practical. Record whether a mismatch was detectable, how long it took to investigate and which fields were missing. Do not reduce the assessment to a match percentage that treats an unsafe false match the same as a harmless unresolved row. The cost of those outcomes differs.
Ask a second operator to explain one accepted mapping from the saved evidence. If they cannot tell which variant is connected or why, the record needs improvement. This handover test often reveals unclear labels that the person who created the mapping no longer notices.
Choose the scope that fits your business
A spreadsheet can be workable when the catalogue is small, sources are stable and responsibility is clear. It becomes harder to maintain when several people edit relationships, variants change frequently or stock decisions depend on current source evidence. Evaluate software around those pressures rather than assuming that catalogue size alone determines the answer.
A supplier-monitoring workflow may fit when you need source observations linked to eBay variations. A broader inventory or warehouse system may be more appropriate when purchasing, locations, kits and order allocation dominate the requirement. These product categories overlap, but their names do not prove equivalent capabilities.
For related buying criteria, read the supplier stock monitoring guide and the eBay inventory software guide. Bring your mapping examples to a discussion with Zelluvo if you want to check the documented supplier-to-eBay workflow against your catalogue. Confirm current source support and account limits before purchase.
Questions about SKU mapping software
Do my supplier and eBay SKUs have to be identical?
No. A mapping can connect different identifiers when the relationship is verified. Preserve both original values and enough source, account and variant context to explain the connection. Identical text alone does not prove that two records describe the same item.
Can product titles be used for matching?
They can help suggest candidates, but titles may omit size, colour, pack quantity or model differences. Confirm a proposed relationship using more precise identifiers and attributes before relying on it for stock decisions.
What should happen when two records have the same SKU?
Check whether the code is unique within the relevant supplier, account or catalogue. If the relationship remains ambiguous, leave it unresolved. Do not let the tool silently choose a destination merely to complete the import.
Does correcting a mapping also correct a live listing?
Not necessarily. A mapping change, a recalculated proposal and a live update are separate actions. Inspect dependent work, approve any appropriate action and confirm the destination rather than assuming the correction propagated automatically.